1 Nonlinearity, Economic Fundamentals and the Dollar-Sterling Real Exchange Rate

With recently proposed econometric methods for summarizing very large macroeconomic data sets into a small number of observable factors, the factor-based fundamental are incorporated into the nonlinear framework in order to investigate the dollar-sterling real exchange rate. We have shown that the information in large data sets is relevant for resolving the purchasing power parity puzzle of international finance and the large information contained nonlinear model can improve the out-of-sample forecasting performance. The accuracy of predictability is examined by the statistical significance of Clark and West (2006) approach, which adjusts the downward bias of difference of error terms in nested models, as well as Diebold and Mariano statistic. These results have important implications for the puzzling aspects of the existing literature on the behaviour of real exchange rates as they suggest that the finding of slower mean reversion and poor out-of-sample forecasting performance may result from a failure properly to allow for the wide range of macroeconomic influences on the equilibrium real exchange rate